Probabilistic Graphical Models

Mission

Comparison of probabilistic graphical models: evaluation of directed Bayesian approaches versus undirected (Markov) models. Advanced benchmark: implementation and comparison of Bayesian networks and Markov networks for cyber threat modeling. Strategic shift: choice of undirected models to explore their flexibility in complex cybersecurity scenarios. Dynamic exploration: in-depth analysis of Markov networks to detect, anticipate and model emerging cyber risks.

Contexte & Environnement

Applied academic research: at the intersection of security, machine learning and advanced statistics. Cybersecurity challenges: threat anticipation, improved detection and system resilience. Heterogeneous data volumes: network logs, incidents, security alert histories. Data and experimentation: dataset preparation, modeling and evaluation of cybersecurity scenarios.

Objectifs Cles

Technologies & Infrastructure